The GUIDE: Magic and the Brain - Part 1 - Illusion

http://teddysratlab.blogspot.com [Full link to blog for email clients.]

At the last "Neuroscience and a Movie," the discussion turned to future programs, and someone mentioned "Neuroscience and Magic."  Several of my colleagues are amateur magicians, and one clinician that I co-taught with last year is actually quite a successful magician.  I have dabbled in the arcane art myself, and while many would term me a "mechanic" (I prefer tricks that rely on a device or physical prop) I do know and understand how a lot of slight-of-hand and illusion work.

Rather than speak directly about magic tricks, I want to devote a few blogs to the two most common features of magic - illusion and misdirection - and how they each play on features of our brain and nervous system. Today I will discuss illusions (and in particular, optical illusions) and the next blog will cover misdirection. [Today's blog is an expansion on a mailbag question from last year.]

Illusions, and in particular optical illusions, are usually caused by one of two processes.  The first is to simply confuse the eyes by playing tricks with what we have come to learn is "normal."  For example, in typical 3-D vision, left is closer to the left eye, right is closer to the right eye, close is big, and far away is small.  A number of the Escher optical illusions take advantage of violating visual rules and conventions.  We can easily follow the lines of the Penrose staircase but the artist violates the rules of perspective by using the same technique of perspective for up/down and near/far.  Likewise the Penrose triangle on the same page violates logic, because instead of consistently shading one surface, Lionel and Roger introduce discontinuities that cannot co-exist, thus creating the illusion.

The second method is to tease the eyes by taking advantage of how the retinal ganglion neurons, lateral geniculate nucleus and V1 visual cortex process vision.  Figure 1 show the distinction between the real world, and the V1 representation.  Because the RGN and LGN are tuned to detect edges, the "fill" in the middle of the text is not represented in V1.  That information is not lost, however, color and fill information is transmitted to V2 and V3 second visual areas, which detect shadings, colorations and start to interpret perspective and parallax. 

Figure 1

When viewed simply as independent lines, the elements of the Vase optical illusion look distinctly like two faces, or a vase (Top of Figure 2).  V1 has no problem distinctly identifying either feature when presented independently.  However, once the lines are put together and shaded, there is a conflict between what V2&V3 say is present (the vase) and what V1 detects (the faces, due to the distinctiveness of the edges of the dark shading).  Thus this second type of optical illusion, like the Necker Cubes (do an internet search, there are way more examples than I can show here).   Rely on the brain being presented with two different interpretations – simply because the visual system processes lines and shading separately.

Figure 2

What about illusions that appear to move, such as in Figure 3?  The "bicycle wheel" illusion at left takes advantage of the fact that the brain and visual system detect motion when successive "line" responsive neurons in the visual cortex are activated.  If one set of neurons, representing a straight line on the left visual field are activated, then a set toward the middle of the visual field, and so on, crossing our vision, the brain interprets the successive activation as indicating motion.  The "ocular dominance" columns of V1 visual cortex consist of groups of neurons that respond to lines of different angles and positions in our visual field.  If the neurons in a given row (i.e. different angles) are activated, we perceive that as rotation of the object - however, in the illusion to the left in Figure 3, we are simultaneously activating the different angle-sensitive neurons, and as our focus shifts to different portions of the figure, the brain interprets the shift as rotational motion of the object.  The parallel lines illusions in Figure 3 again rely on the representation of lines in V1.  Our perception is biased to look for straight lines, and curves are really only detected by difference between the curved and straight line.  Thus the vertical pair of lines at the center of each cluster of four lines is identical, and perfectly parallel, but the differing angles and spacing produced by the curved lines tends to fool the brain into think that all of the lines are curved.

Figure 3

One of the more interesting "psychometric" aspects of neuroscience is that it is possible to detect when a person's perception of an optical illusion shifts.  Most of the motor control are of the brain is in the frontal cortex, just forward of the border with the parietal cortex, and control of the eye muscles is now exception.  It may seem that occulomotor control (Cranial Nerve III, the third "O" in yesterday's mnemonic) is a simple matter of pointing the eyes in the right direction.  However, the process is *much* more complicated, requiring actual target acquisition and identification – in other words, the full suite of visual cortical processing.   Distance and horizontal tracking requires that the eyes move at slightly different angles; focus and lighting changes requires pupil diameter control.  The "Frontal Eye Fields" along (with the Edinger-Westphal nucleus of deep thalamus and the superior colliculi and locus coeruleus of the brainstem) is involved in the complex process of integrating actual vision with the process of adapting the eye to light and motion.   When visual information changes it can be revealed as changes in scanning the environment or reacting to light. 

In fact, it has been demonstrated that if a person is shown a Necker Cube-style optical illusion, and told to press a button whenever their perception of the cubes changes from the "top" to the "bottom" view, the pupils dilate briefly.  This is just one small way in which the operation of the brain (or – dare I say it – The Mind) can be monitored by a physiological reaction. 

Figure 4

One last type of optical illusion that depends on even more complex association of vision and language, and it gets to the types of illusions that magicians employ in their acts:  The "Stroop Interference" effect shown in Figure 4 violates the consistency of line vs. shading, but also introduces understanding of the word meaning.  This process depends heavily on the multi-sensory association cortices at the intersection of Occipital, Temporal and Parietal lobes – with the added involvement of decision making by the Frontal Lobe.  This is one of those phenomena that belies the idea that we only tap a tenth of our brain.

So, how does a magician fool our eyes and our brains?  Rather simple, really, the illusionist knows that the brain is looking for fairly simple features, and that combining vision with other information (such as sound, language, or memory) can cause conflicts between what we think we see, and what is really there.  We perceive closed lines and circles when they are not, or openings where there are none.  We confuse language with color with angle and notice the familiar among  the unusual.  A large percentage of the magician's craft is also misdirection, and we will discuss that in the next blog.

Until then, don't confuse your brain, take care of it, and watch out for illusions!